Economic Compass
ISSUE 37 • JANUARY 2016
GLOBAL PERSPECTIVE FOR INVESTORS
THE FUTURE OF PRODUCTIVITY AND INNOVATION
Eric Lascelles
Chief Economist
HIGHLIGHTS
nn
Global productivity growth has
slowed since the financial crisis,
worrying many.
nn
Much of this productivity
deceleration represents an
inevitable slippage after a period
of unusually rapid gains. But some
also reflects temporary cyclical
depressants that have taken hold
since the crisis.
The world is embroiled in a crucial debate about the future of
productivity. Simply put, is it set to falter, hold steady or instead
accelerate (Exhibit 1)? The importance of this question could
not be greater since, over the long run, rising productivity and
the innovation that drives it are the keys to economic prosperity.
Raising the stakes, the economic outlook is already clouded by
demographic challenges. The performance of productivity and
innovation will thus largely determine whether economic growth is
decent or dismal.
There are strong arguments on both sides of the productivity debate. Pessimists
dwell on a series of fading one-time productivity dividends that came from
urbanization, universal education and the widespread entry of women into the
workforce. Similarly, as emerging economies become wealthier and approach the
technological frontier, they may find less scope to grow through the absorption of
technologies from the developed world.
However, there are no less compelling claims from the optimists. Basic science
continues to advance at a heady clip, a raft of new technologies is dramatically
disrupting an expanding set of industries, and emerging-market economies should
become increasingly capable of driving innovation themselves rather than simply
importing it.
nn
Looking forward, productivity growth
should manage a gradual revival as
the rate of innovation accelerates
across a range of sectors.
Exhibit 1: U.S. productivity waves
nn
However, unusually fast productivity
nn
As part of this exploration, we
evaluate new technologies and their
potential effects by sector.
nn
We also consider whether
productivity growth is being
mismeasured, whether the world
is shifting toward a “capital light”
economy, and whether there might
be significantly more structural
unemployment in an increasingly
automated world.
5
U.S. productivity growth
(annual % change, 10-year average)
growth is unlikely given slightly less
help from capital investment, labour
quality and technological diffusion.
Railroad boom
Post-war boom
4
Mid-war boom
I.T. boom
3
2
?
1
0
-1
-2
1899
1922
1945
1968
1991
2014
Note: Data prior to 1948 is consumption per capita growth; 1948 and later is standard output per hour.
Source: Haver Analytics, R. Shiller, RBC GAM
In weighing the various perspectives (Exhibit 2), we
ultimately conclude that the current productivity slowdown
is mostly temporary, with additional innovation set to fully
compensate for fading tailwinds, weaker technological
diffusion and a less forceful contribution from capital
deepening and labour quality.
As part of this exploration, we evaluate new technologies and
their potential effects by sector. We also consider whether
productivity growth is being mismeasured, whether the world
is shifting toward a “capital light” economy, and whether
there might be significantly higher structural unemployment
in an increasingly automated world.
Where productivity growth comes from
The only way to wring ever more out of a fixed amount of
human effort is by becoming more productive. Productivity
growth is the key to a rising standard of living, robust
economic growth and – ultimately – financial-market returns.
Fortunately, there are quite a number of ways to improve
productivity (Exhibit 3).
The simplest, most immediate and most predictable means
of increasing productivity is to increase a country’s or
company’s capital intensity. That means providing more
capital such as machinery, equipment and structures to
boost worker output. Over the past 20-plus years, increased
capital intensity has generated a whopping 60% of
developed-world productivity growth (Exhibit 4).
Another reliable strategy – if less powerful and slower to pay
dividends – is to increase the quality of the labour force via
education and training. Rising labour quality has historically
generated 12% of productivity gains.
Lastly and of crucial importance to productivity growth is
the delivery of rising total-factor productivity (TFP). This
amounts to finding innovative ways to more efficiently deploy
a preset amount of capital and labour. This has officially
generated 28% of productivity gains in the modern era,
but its true importance is even higher: productivity growth
would eventually grind to a complete halt in the absence of
new technologies and ideas. Examples of TFP gains help to
illustrate its fundamental importance:
nnAt its most exciting, TFP can rise due to brand new
technologies that propel the economy forward. This is
innovation in its purest form, and encompasses the
development of inventions both large (electricity, the
computer) and small (the paperclip).
2
|
Economic Compass • Issue 37
Exhibit 2: Productivity pros and cons
Productivity accelerants
Productivity decelerants
Fading 20th century tailwinds
Innovation
Capital investment
Labour quality
Eventual end of post-crisis malaise
Tech diffusion
Net effect: normal productivity growth
Source: RBC GAM
Exhibit 3: Where does productivity come from?
Productivity
growth
Capital
intensification
Rising labour
quality
Total factor
productivity
growth
Diffusion of
existing innovation
Innovation
Better
technologies
Better
processes
Brand new
innovations
Incremental
improvements
Diffusion
across
countries
Diffusion
across
companies
Diffusion
across
people
Source: RBC GAM
Exhibit 4: Historical drivers of developed-world
productivity growth
TFP
28%
Capital
investment
60%
Labour
quality
12%
Note: Productivity contribution share for thirteen major developed countries.
3-year moving average between 1992 and 2007. TFP is total-factor
productivity. Source: The Conference Board Total Economy Database, Haver
Analytics, RBC GAM
Economic Compass
nnSlightly less exciting but far more common is innovation in
the form of incremental technological improvements – the
latest computer model, for instance.
nnInnovation does not merely occur via the invention of
physical things. A better process can be just as important.
These improvements can be large (a rejigged supply chain),
middling (an improved sequence of tasks on a factory floor)
or small (eliminating a bit of wasted fabric).
nnPossibly the least acknowledged and yet most important
driver of global TFP growth is the diffusion of pre-existing
technologies into new parts of the world and more widely
across companies. Only a handful of the world’s richest
nations operate at the technological frontier, meaning
that they are reliably developing new technologies. For
the rest, their technology improves mainly by absorbing
and implementing ideas that have already been invented
elsewhere.
(Appendix A), and also that human well-being may be rising
more quickly than even the recalibrated economic figures
suggest (Appendix B).
Nevertheless, even factoring in these adjustments, it seems
likely that true productivity growth has declined. The next
logical question is whether this productivity deceleration
is temporary or structural, as that will determine whether
households and investors are merely being inconvenienced
or permanently damaged.
Temporary influences
There are a number of temporary distortions to productivity
growth due to past cyclical tailwinds that have faded, and
some newer headwinds that have temporarily strengthened.
Prior temporary tailwinds
Of course, these three types of productivity growth are all
fundamentally intertwined. Educated humans invent new
technologies, which in turn enable an ever-evolving capital
stock.
Prominently, it is worth acknowledging that productivity
growth from the mid-1990s to the mid-2000s was unusually
good (Exhibit 6). Thus, a substantial chunk of the subsequent
decline in productivity growth is simply a reversion to a more
normal trend rate after an unusually strong prior cycle.1
Recent slowdown
Future temporary headwinds
Given the importance of productivity growth, it is distressing
that the rate of ascent has tumbled quite abruptly in recent
years in both developed and emerging economies. The
decline in developed-world productivity growth since the turn
of the millennium has been particularly broadly based, with
less assistance coming from all three drivers (Exhibit 5).
Meanwhile, the global financial crisis casts a cyclical
shadow over productivity growth today. This means that the
recent dismal productivity performance is in significant part
temporary and should begin to improve as time passes.
We take some solace in our assessment that GDP and thus
productivity growth may be somewhat underestimated
Exhibit 5: Developed-world productivity slows for multiple
reasons
Even as risk appetite revives and business investment has
picked up, the productivity hangover is likely to persist for
1
This boom was largely the result of the initial impulse of the information
technology revolution, and also partially due to unsustainably fast credit growth
and questionable financial sector deregulation.
Exhibit 6: U.S. productivity growth was abnormally
strong beforehand
3.2
U.S. business-sector real output per
hour (10-year % change annualized)
Contribution to GDP growth (ppt)
3
2
1
0
-1
1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014
Total factor productivity (TFP)
Labour quality
Capital
Note: 3-year moving average of contribution to GDP growth of thirteen major
developed countries. Source: The Conference Board Total Economy Database,
Haver Analytics, RBC GAM
Unusual
productivity
spike
2.8
2.4
Reversion
to more
normal
growth...
2.0
... and
moderate
overshoot
1.6
1.2
0.8
1980
1985
1990
1995
2000
2005
2010
2015
Note: Real output per hour of all persons in business sector.
Source: BLS, Haver Analytics, RBC GAM
| 3
somewhat longer. As the economy undershot its potential
for eight years, business investment was persistently weaker
than normal. This leaves the capital stock smaller than it
would have been – a key determinant for productivity growth.
Thus, even as business investment revives and the economy
nears its full capacity, the lower level of capital intensity will
persist for many years. Furthermore, IMF research finds that
a downward shock to capital investment also tends to have a
negative effect on the rate of innovation (TFP growth).
Exhibit 7: Urbanization continues rising, but nearing its
limit?
OECD urban population
(% of total population)
85
There is a similar cyclical hole in the labour market. Labour
quality is unquestionably lower than it otherwise would
have been. Many millions of workers were unemployed
or underemployed for an unusually long period after the
financial crisis. As a result, their existing skills decayed and
they failed to acquire new ones. This hole will also take many
years to plug.
A further side-effect of the financial crisis is that interest
rates have been extraordinarily low for many years. Low
interest rates are sometimes thought to gradually unleash
two negative influences on productivity growth. First,
the marketplace can become cluttered by zombie firms
that would have been shuttered in a normal interest-rate
environment. Second, and relatedly, low rates reduce the
threshold for investing, resulting in projects being pursued
that wouldn’t be deemed worthwhile under more normal
circumstances. Once interest rates normalize, businesses
and investors regret these low-yielding projects.
The broad conclusion is that productivity growth enjoyed an
unusually strong cyclical upswing through the late 1990s
and early 2000s, and so it should not be surprising that this
has come to an end. Simultaneously, there are some specific
headwinds related to the financial crisis that will linger for
several more years, but are not permanent impediments.
Thus, the cyclical productivity trend is near its trough and can
be expected to improve somewhat over time.
Structural influences
These cyclical observations are promising, but far more
important is the question of whether the world’s capacity
for productivity growth is rising or falling in a structural (i.e.
semi-permanent) way.
The U.S. Federal Reserve’s (Fed) productivity model argues
that we are still in a low productivity regime – in other words,
80
75
70
65
60
55
50
1960
1966
1972
1978
1984
1990
1996
2002
2008
2014
Source: Haver Analytics, World Bank, RBC GAM
that this is a structural experience rather than merely a
cyclical one. That said, the model’s definition of “low” versus
“high” productivity growth is not entirely in line with our own
thinking. Its definition of high is a searing 3.0% productivity
growth. We would be delighted with sustainable productivity
growth in the range of 1.75% to 2.0%, never mind 3.0%. In
fact, the Fed’s model predicts precisely this over the next
five years.
The case for less productivity growth in the future
The standard argument for less productivity growth in the
future is that most of the easy pickings that enabled such
rapid productivity growth in the past are now becoming
scarcer.
Urbanization: The widespread migration from farms to cities
has given productivity a critical boost. Each percentage-point
increase in the rate of developed-world urbanization adds
0.75% to the level of productivity. In other words, newly
urbanized workers are nearly twice as productive as when
they were on the farm. In emerging markets like China, the
rural-to-urban multiplier is more like three times. However,
now that the majority of developed-world citizens live in
urban agglomerations, the scope for further support from this
trend seems destined to fade in the future (Exhibit 7).2
Female employment: The widespread entry of women into
the workforce over the past few generations has provided
an enormous boost to economic output, and a smaller
2
Providing a further indignity to overall economic growth, now that the
urbanization trend is largely completed in the developed world, one of
the realities of a highly urban world is a lower fertility rate. This limits the
demographic contribution to economic growth as well.
4
|
Economic Compass • Issue 37
Economic Compass
boost to productivity.3 But the upward trend in the female
employment rate has since halted, meaning this structural
tailwind is gone (Exhibit 8).
Exhibit 8: Female boost to economic and productivity
growth
U.S. employment-population ratio (%)
90
Education: The quality of the world’s human capital has
increased enormously for several centuries, starting with
an expansion of universal access to education, followed
by rising high school graduation rates and lately spiking
university attendance. However, it seems obvious that not
everyone will have a PhD in 50 years, nor would this even be
optimal given that years of potential work must be sacrificed
in such a pursuit. Thus, the remarkable rise in education
standards will struggle to continue advancing at the same
rate into the future.
Longevity: Although longevity continues to rise, the pace is
decelerating, and the extra years no longer secure a material
increase in workers, productivity or all-around quality of life.
Source: BLS, Haver Analytics, RBC GAM
Land: At least in the “New World,” land was effectively free in
the early stages of colonization, eliminating a key capital cost
for businesses. But as the population and economy have
grown, the remaining space has become progressively less
desirable and more expensive. Companies must now spend
far more money on the same underlying capital stock.
Entitlements: Many government entitlements were created
on a pay-as-you-go basis with the presumption that steady
population growth would continue indefinitely. This
assumption has proven incorrect. Now that fertility rates
have declined and public-debt levels have risen, these
expenses will grow as a share of government revenue, pulling
government resources away from more productive ends, such
as infrastructure spending or lower tax rates.
A further popular claim is that the past few centuries were
blessed by the serendipitous discovery of a raft of profound
new technologies, each of which lifted productivity growth
materially for several decades. Such “general purpose
technologies” include the railroad, electricity, the telephone,
mass production, the combustion engine, the corporation,
the radio, the air conditioner, plastics and antibiotics. Each
had enormous productivity-enhancing effects across a wide
swath of human life, and laid the foundation for additional
innovations. The presumption is that there will be fewer
inventions of this scale in the future.
80
70
60
50
Gap between male and female
employment rate
has narrowed over time...
40
...but the female
employment rate has
since flattened
30
20
1948
1959
Male
1970
1981
1992
2003
2014
Female
for innovation increased dramatically at the onset of the
Industrial Revolution. We have not detected any compelling
evidence that this remarkable capacity for invention has
suddenly been snuffed out. Axiomatically, there should
continue to be major new discoveries in the future. We
can think of several that are unfolding right now, such as
computers, networks and robotics. Others will come as a
pleasant surprise, much as the course of prior discoveries
was hardly mapped out in advance.
But we get ahead of ourselves. To offer a comprehensive
opinion on the outlook for productivity, we must evaluate
each of the determinants of productivity highlighted in
Exhibit 3 to gauge their adequacy in combatting the loss of
structural tailwinds just identified.
Normal current capital contribution
The first and traditionally largest driver of productivity growth
is via a rising capital stock. Although the stock of capital is
undoubtedly smaller than it would have been without the
financial crisis, it has nevertheless managed to roughly keep
pace with GDP (Exhibit 9). This argues – theoretically – for a
fairly normal contribution to productivity growth, if less than
what might have been without the financial crisis.
We cannot deny the tremendous discoveries of the past, but
reject the assumption that the future will be any less exciting.
For a variety of complicated reasons, the world’s capacity
Sending a positive signal about the next few years, the rate of
U.S. private-sector capital investment growth is rising and is
actually slightly above normal on an inflation-adjusted basis
(Exhibit 10).4 Technically, this is even more promising than
it first seems as the level of capital investment necessary
to sustain a fixed level of capital intensity falls with a lower
3
4
This has effectively doubled the number of people with a high level of human
capital in the economy.
Capital goods prices have generally risen less quickly then other products,
making nominal capital investment look paltry even when the volume of new
capital purchases is rising.
| 5
Naysayers like to point out that businesses are currently
deploying an unusually small fraction of their profits into
capital investment (Exhibit 11). This is true, but largely
because profits are unusually high rather than business
investment being unusually low. Meanwhile, if firms were
to become more enthusiastic about the future – or if the
market’s preference for dividends and buybacks were to
5
This explains in significant part why China must dedicate so much of its output
to capital investment: otherwise, its rapid GDP growth would quickly outstrip its
capital stock and limit future economic gains.
The composition of capital investment has also improved
as firms have increasingly tilted their purchases toward
especially productivity-enhancing capital such as information
and communication technology (ICT) and intellectual
property. Intellectual property now represents fully onequarter of capital investment (Exhibit 12). On the other hand,
these types of capital depreciate unusually quickly and so
require constant replenishment.
Thus, our initial prognosis is that the current pattern of
capital investment is consistent with a slightly larger than
normal contribution to productivity growth. But this is not the
entirety of the story, as there is considerable debate about
whether the changing contours of the economy might make
Exhibit 10: Business investment rising again
450
22
400
20
Nominal business investment
(% of nominal GDP)
Real capital stock
(% of real GDP)
Exhibit 9: Capital intensity has held up well to recent crisis
fade – they could easily afford to increase the pace of
investment further.
350
300
250
200
150
100
Japan
1970s
Germany
1990s
U.S.
China
16
14
14
12
12
10
1987
1994
2001
Nominal Business Investment (LHS)
Latest decade
2008
2015
8
Real Business Investment (RHS)
Note: Use of different scales to align historical averages.
Source: BEA, Haver Analytics, RBC GAM
Exhibit 11: U.S. firms allocate less of profits to capex
Exhibit 12: Rising intellectual property investment
28
Investment could
be higher
320
240
160
80
1957
1967
1977
1987
Source: NBER, BEA, Haver Analytics, RBC GAM
Economic Compass • Issue 37
1997
2007
2017
U.S. intellectual property investment
(% of private fixed asset investment)
U.S. nonresidential fixed investment
(% of corporate profits)
400
0
1947
|
16
Note: Average real capital stock as % of GDP for the periods indicated. Latest
decade from 2002 to 2011. Source: Feenstra, Robert C., Robert Inklaar and
Marcel P. Timmer (2015), “The Next Generation of the Penn World Table”
forthcoming in American Economic Review, available for download at www.
ggdc.net/pwt; RBC GAM
480
6
18
Historical
average
18
10
1980
India
20
Real business investment
(% of real GDP)
economic speed limit.5 Developed world growth rates are
lower than before due to deteriorating demographics. Thus,
if the current enthusiasm for business investment persists, a
notably rising capital intensity is conceivable.
Intellectual property
rising as a share of
capital investment
24
20
16
12
8
4
1954
1964
1974
1984
1994
2004
2014
Note: Nominal fixed asset investment. Source: BEA, Haver Analytics, RBC GAM
Economic Compass
capital investment a lower priority for companies in
the future.
Exhibit 13: U.S. consumers embrace retail e-commerce
Slightly less capital contribution in the future?
8
The consensus thinking is that businesses may be starting to
transition toward a “capital light” world as new technologies
and technology-oriented companies displace their brickand-mortar ancestors. E-commerce sales are making steady
inroads into retail spending, rising from almost nothing
in 1999 to 7.4% of sales today, and continue to capture
additional market share (Exhibit 13).
U.S. retail e-commerce sales
(% of retail sales)
presents precisely the opposite argument: the expected
increased usage of robots, machines and software all argue
for considerable future capital investments.
nnWith reference to the shift toward the service sector, a
2
2003
2006
2009
2012
2015
nnAs this disruption to “old economy” companies occurs, a
fierce battle is erupting over who will capture the market
share that is suddenly up for grabs. This competition
may manifest in the form of accelerating investment and
corporate acquisitions (another form of capital investment).
nnThe capital stock has not yet shown any inclination to shrink
nnRelated to this, while a handful of sectors – most notably
social networks and search engines – are enjoying greater
barriers to entry thanks to unstoppable network effects,7
the bulk of sectors – such as media, entertainment and
trade – are experiencing reduced barriers to entry. The
enhanced competition that results may create a sort of arms
race that incents additional capital investment.
nnThe definition of capital is broader than conventionally
storefronts, keep in mind that the sale of goods demands
an enhanced wholesaling infrastructure (larger, more
sophisticated warehouses) and obliges far more
transportation capital as products are shipped directly to
homes rather than en masse to stores. These will provide a
material offset to reduced storefronts.
3
capture the bulk of all retail spending, it is not reasonable to
assume that store-based transactions will vanish altogether.
There are inherent limits to how much can be purchased
online given the value that some place in the social and
experiential aspects of shopping, personalized expert
advice from salesclerks and the ability to physically evaluate
a product before purchase.6
nnAnother powerful trend on the upswing – automation –
nnEven in a world of online sales that decimates physical
4
Source: U.S. Census Bureau, Haver Analytics, RBC GAM
However, we believe the trend toward “capital light”
operations will prove much less powerful than imagined:
imagined, as it includes forms of intellectual property such
as software, research & development and patents. The
acquisition of such soft capital should accelerate given
the anticipated technological focus. Meanwhile, these
new kinds of capital are particularly good at driving
productivity gains.
5
0
2000
“Capital light” counterpoints
even as profound new technologies have unfurled over the
past 15 years.
6
1
Some also argue that the broader economic shift away from
goods and toward services should further decrease capital
intensity.
surprising fraction of service industries are in fact very
capital intensive – such as rail transportation, telecom and
healthcare. This transition does not automatically result in
less capital intensity.
Still a small share but
rising briskly
7
Return on capital
Framed slightly differently, so long as a healthy return
on capital persists, why would companies stop investing
in capital? The current U.S. return on capital is actually
unusually high, suggesting that the substantial technological
changes already under way have yet to seriously undermine
6
nnWhile e-commerce sales should continue to expand their
market share for decades to come and may ultimately
Offering free returns for delivered products is not the same, as it is a hassle to
repackage and return them.
7
Absent significant differentiation, no one will join a new social network unless
all of their friends are already there, making it an almost impossible task to
compete with successful incumbents.
| 7
Even if the demand for capital ebbs slightly in the future –
theoretically exerting a downward force on the marginal
return on capital – keep in mind that the dissaving that
accompanies an aging population should reduce the supply
of investible funds in parallel (Exhibit 15). These two forces
should roughly neutralize each other, leaving the ultimate
return on capital essentially unchanged.
As such, we walk away with the impression that while
“capital light” expectations are not entirely an illusion, the
effect may not be nearly as strong as most expect. Thus,
we budget for a bit less productivity help from capital
investment, but only a bit.
Exhibit 14: High return on capital is good sign for future
growth
High return argues for
more capital investment
120
Return on capital
(Index Q1 1980=100)
the value of investing in capital (Exhibit 14).8 Companies are
still well incented to invest.
Historical
average
100
80
60
40
20
0
1980
1987
1994
2001
2008
2015
Source: BEA, Haver Analytics, RBC GAM
Labour quality contribution
nnActual measures of global education continue to rise
without any apparent sign of deceleration. In fact, if
anything, there has been a slight acceleration (Exhibit 16).
Much of this success comes from emerging-market nations.
nnThe “Flynn effect” of rising IQ scores from generation to
generation – theoretically independent of education –
continues to add to labour quality in a somewhat mysterious
way, possibly related to rising levels of stimulation in the
world and/or improving nutrition standards.
Exhibit 15: Dissaving could keep return on capital from
falling
30
% of total population
As discussed earlier, labour quality has been an important
driver to productivity growth for centuries. There are
legitimate fears that this upward trend could be coming to
an end. We share this concern for the developed world, but
should note that there are several countervailing factors from
a global perspective:
24
Those under 45 are
roughly neutral
18
12
Faster saving growth than
spending growth
6
1980
1990
2000
Age 45 to 64
2010
Savers shrink / dissavers
grow
2020
2030
2040
2050
Age 65 and over
Note: Population of developed countries. Source: United Nations, Haver
Analytics, RBC GAM
nnLabour quality may also be able to continue rising in a more
Overall, we conservatively presume that labour quality will
provide slightly less help than in the past to productivity
growth, but it may not fade as much as many imagine.
Exhibit 16: Education quality not done rising
35
World enrollment in tertiary education
( % of total tertiary-age population)
unorthodox fashion as increased flexible work arrangements
save commuting time, reduce stress and allow for a more
efficient distribution of work across the day.9 Similarly,
increasingly diverse workforces may unleash positive social
dynamics with favourable consequences for productivity.
30
25
20
15
10
5
8
Of course, there is a slight circularity to this argument given that this measure
is calculated as profits as a percentage of the corporate capital stock, such that
a marginal decline in capital actually increases the return on capital, at least
initially.
9
Of course, these advantages must be contrasted against potential
disadvantages, such as reduced communication across teams and the scope for
distraction at home.
8
|
Economic Compass • Issue 37
0
1970
1975
1980
1985
1990
1995
2000
2005
2010
Note: Tertiary education includes all post-secondary education, including but
not limited to universities. Total school-age population is the total population
of the five-year age group following on from secondary school leaving.
Source: World Bank, Haver Analytics, RBC GAM
Economic Compass
Technological contribution
Exhibit 17: Tech diffusion vs. tech innovation
With the prospect of slightly less support from both capital
intensification and labour-quality growth over the long run,
any hope of achieving a normal rate of productivity growth
lies in the hands of TFP growth, and whether it can revive
after having shrunk to virtually nothing in recent years (refer
back to Exhibit 5 for developed economies and forward to
Exhibit 18 for emerging economies).
Tech diffusion
20 %
Description:
Current:
We believe there could be less technological diffusion in the
future, but more outright technological innovation (Exhibit
17). On the aggregate, the latter should outweigh the former.
Developed
Tech innovation
60%
Emerging
80 %
Less
Future:
Emerging
More tech innovation
from EM as they
approach tech frontier
Less tech diffusion
from rich nations as
EM become richer
Expected
change:
40%
Developed
More
Less technological diffusion
Source: RBC GAM
A key driver of worldwide productivity over the last century
was the better diffusion of technology from rich nations to
poor ones. This enhanced flow occurred for a number of
reasons. Extreme differences in technological advancement
created the opportunity. Rising trade, direct investment
flows and a global diaspora then brought the technological
advancements to poor nations. Finally, many poor countries
proved increasingly capable of absorbing and re-deploying
these technologies when they encountered them.10
Exhibit 18: EM productivity growth slows for one key
reason
Contribution to GDP growth (ppt)
9
Whereas technological diffusion has generated just 20% of
TFP growth for rich nations, it has provided a remarkable 60%
of the gain for poorer nations.
In the future, we suspect technological diffusion will weaken
slightly, for the following reasons:
nnTrade and foreign direct investment are acknowledged to be
key drivers of technological diffusion. We believe the pace
6
5
4
3
2
1
0
2002
2004
2006
Total factor productivity (TFP)
2008
2010
Labour quality
2012
2014
Capital
Note: 3-year moving average of contribution to GDP growth of six major EM
countries. Source: The Conference Board Total Economy Database, Haver
Analytics, RBC GAM
of globalization is slowing in a structural way, which may
interfere with this process.11
nnSome companies on the bleeding edge of technology have
stopped securing patents for their findings to avoid allowing
competitors to see the exact specifications of their new
inventions.12 This may limit cross-firm diffusion.
nnAs emerging-market nations become wealthier, they are
naturally approaching the technological frontier. This
reduces the scope for further technological absorption.
Indeed, emerging-market TFP growth has plummeted
(Exhibit 18). Recall that 60% of this has traditionally come
from technological diffusion.
7
-1
nnBetween nations, income inequality has been declining.
Similarly, countries are becoming increasingly homogenous
in terms of their education levels, business activities,
cultures and tastes. No longer do the shopping malls
of foreign countries reveal a kaleidoscope of unfamiliar
products. They all look and operate in an increasingly similar
fashion, meaning less remaining room for the diffusion of
unfamiliar technologies from one nation to the next.
Lost TFP growth
explains all of
productivity
slowdown
8
To be fair, it is hardly a one-way street of declining technology
diffusion. The internet makes the spread of basic knowledge
and technologies much easier. In addition, immigrants
regularly take technological know-how back to their country
of origin. Finally, decent labour-market mobility allows for the
11
10
Countries were more capable of absorbing foreign technologies due to greater
economic flexibility that rewarded ideas more generally.
Trade flows are already weak. While FDI flows are still strong, companies
seem decreasingly willing to effectively give away their technological secrets to
partners in foreign countries.
12
On the other hand, cyberattacks and open-source software are laying bare
technologies that might otherwise have gone unshared.
| 9
diffusion of knowledge technologies across firms – another
important channel.
Despite these counterpoints, we believe the balance
of evidence favours an interpretation of slightly less
technological diffusion at the global level.
More innovation
This takes us to the crux of the productivity debate: will
the rate of innovation rise or fall in the future? With plenty
of caveats, we are more inclined toward the optimistic
argument.
Leading indicators
Before we evaluate the promise of specific new technologies,
let us consider the signals being sent more broadly. There are
quite a number of positive ones:
Exhibit 19: The long road to productivity gains
Basic
research
Disruptive
innovation
Years/
Decades
Years
Decades
EXCITEMENT:
Unexciting
Exciting
Less
exciting
ECONOMIC
VALUE:
No benefit
Small
benefit
Maximum
benefit
TIMING:
Deployment
Source: Carlota Perez, Technological University of Tallinn, RBC GAM
Exhibit 20: Middle-income country innovation on the rise
nnExisting “general purpose technologies” – revolutions like
the computer and the network – continue to play out, as
we discuss in more detail later. Historically, monumental
achievements like these take decades to be fully
absorbed into productivity figures, and they enable other
technological improvements on their backs (Exhibit 19).
These technologies can continue to pay dividends.
nnThere are fewer barriers to innovation today than ever:
basic knowledge is available to anyone on the internet;
many sectors are now ripe for disruption by startups; the
venture capital industry continues to expand and improve;
off-the-shelf tools now exist to start a sizeable business
almost overnight via plug-and-play solutions such as selling
on Amazon, advertising on Google and storing data in the
cloud; and large companies are keen to pay up for good
ideas by bolting on small acquisitions.
nnAlthough emerging-market nations may find there are fewer
foreign technologies they do not already possess than in the
past, they should be theoretically better at generating their
own technological innovations now that they are nearing
the technological frontier. This is evident in the data, with
emerging-market R&D spending and patent applications
both increasing sharply (Exhibit 20). As a result, the pool
of prospective innovators has effectively expanded from
less than a billion people in the developed world to many
billions around the world.
10
|
Economic Compass • Issue 37
800
1.4
700
1.2
600
1.0
500
0.8
400
0.6
300
0.4
200
0.2
100
0
R&D expenditure (% of GDP)
is not obviously impaired, nor have all the mysteries been
solved in basic science. Many fields are just scraping the
surface. Every new scientific or technological solution
surfaces new burning questions.
Patent applications (thousands)
nnThe fundamental creativity and inquisitiveness of humans
1985 1988 1991 1994 1997 2000 2003 2006 2009 2012
Patent applications (LHS)
0.0
R&D expenditure (RHS)
Note: Patent applications and R&D expenditures of middle income countries.
Source: World Bank, Haver Analytics, RBC GAM
nnMeanwhile, developed-world R&D is also rising. U.S.
inflation-adjusted R&D has recently accelerated out of a flat
decade and is now at its highest share of GDP on record
(Exhibit 21). R&D spending tends to lead productivity
growth by several years.
nnThe time between technological discoveries and
commercialization, while still long, may be shrinking
(Exhibit 22). This allows future technologies to map much
more quickly onto productivity.
Admittedly, there are also a few unpromising signals, though
we believe they are trumped by the prior factors:
nnOne clever strategy for divining the financial market’s own
opinion on the expected degree of innovation is to look
at the gap between the price-to-earnings ratio of the tech
sector and that of the overall stock market. This reveals a
pretty low premium, and so fails to identify an expected
surge (Exhibit 23).
Economic Compass
nnHistorically, a great deal of innovation comes from “creative
destruction,” whereby new technologies and firms outright
replace old ones. However, the turnover of businesses has
actually been falling for quite some time (Exhibit 24).
New technologies
We now turn to our crystal ball to evaluate the key new
technology buckets likely to drive productivity growth over
the next few decades (Exhibit 25).
In so doing, Amara’s Law is worth heeding: “we tend to
overestimate the effect of a new technology in the short
run and underestimate the effect in the long run.” Thus,
we should withhold judgement even as some of these
technologies get off to a slow start.
Let us also acknowledge that the yardstick for technological
improvement is constantly changing. Once upon a time, a
key test was how quickly one could get from London to New
York. Humankind has failed to advance at all in this regard
in recent decades. However, one can now instantaneously
obtain all of the information from the libraries of London in a
split second, and speak face-to-face with someone there via
video conference. Technological advancements do not always
come in a linear fashion, but they do still come.
Computers
Computers are an indisputable “general purpose technology”
given their widespread application. They have now been
permeating developed economies for the better part
of 50 years, and have had a profound effect on society
and productivity. The current level of data processing
and analysis was not remotely possible when performed
manually by humans.
Exhibit 22: The pace of technological adoption may be
accelerating
Exhibit 21: R&D is rising and leads productivity
50
...now rising again
U.S. real R&D (% of real GDP)
1.80
1.60
1.40
1.20
1.00
R&D share stagnated for 2 decades...
0.80
0.60
0.40
Years until technology adopted by
25% of American population
2.00
40
30
25
20
15
10
1975
1985
1995
2005
1873
1875
1897
Television
26
Personal
computer
16
Mobile phone
13
Web
7
1926
1976
1983
1991
Source: Haver Analytics, RBC GAM
Source: BofA Merrill Lynch, Ray Kurzweil, Singularity.com, RBC GAM
Exhibit 23: Markets not pricing in a sudden surge of
innovation
Exhibit 24: Firm turnover declining steadily
16
2.0
1.5
1.0
0.5
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014
Source: Bloomberg, RBC GAM
U.S. firm turnover (%)
Ratio of S&P 500 Information
Technology P/E to S&P 500 P/E
15
3.0
2.5
2015
First commercially available year
2015
3.5
5
0
1965
Telephone
35
Radio
31
35
0.20
0.00
1955
Electricity
46
45
14
13
12
11
10
9
8
1977
1983
1989
1995
2001
2007
2013
Note: Turnover defined as fraction of outstanding firms created or closed in a
year. Source: U.S. Census Bureau, RBC GAM
| 11
Exhibit 25: Evaluating new technologies
Category
Development stage
Importance
Productivity growth impact
Calculation
Late
High
Accelerate
Pattern recognition
Mid
High
Accelerate
Artificial intelligence
Early
High
Accelerate
Internet
Late
High
Accelerate
Email
Late
High
Sustain
Video conferencing
Mid
Low
Decelerate
Mobile
Mid
High
Accelerate
Virtual reality
Early
Low
Decelerate
Precision
Late
High
Sustain
Manipulation
Mid
High
Accelerate
Miniaturization
Mid
Medium
Decelerate
Social intelligence
Early
Low
Accelerate
Task flexibility
Early
Medium
Accelerate
Driverless car
Mid
High
Sustain
Driverless trucks
Mid
Medium
Sustain
Drones
Mid
High
Accelerate
Materials
New composites
Early
Medium
Decelerate
Energy
Shale oil & gas
Late
Low
Sustain
Wind
Mid
Medium
Sustain
Solar
Mid
High
Sustain
Biofuels
Early
Medium
Sustain
Batteries
Early
High
Accelerate
Farm analytics
Mid
Medium
Sustain
Genetic engineering
Early
Medium
Sustain
Drug development
Late
High
Sustain
Health tracking & diagnosis
Mid
High
Sustain
Robotic surgery
Mid
Low
Sustain
Synthetic biology
Early
High
Accelerate
Big data analytics
Early
Medium
Sustain
Custom manufacturing
Mid
Medium
Accelerate
Niche targeting
Mid
High
Accelerate
3D printing
Early
Medium
Accelerate
Computers
Networks
Robots
Transportation
Food
Health
Customization
Technology
NET EFFECT:
MODERATE ACCELERATION
Note: A decelerating impact on productivity growth means the new technology will still add to productivity, but won’t provide as much upward momentum as prior
technologies in the space did. Source: RBC GAM
12
|
Economic Compass • Issue 37
Economic Compass
Perhaps the most important question for the future is
whether the boost to productivity growth from computers
may now be starting to fade after many decades of support.
After all, some of the most elemental uses for computers –
performing calculations, for example – have already been
widely deployed.
integrated into education, healthcare, banking, agriculture
and elsewhere.
We reject this concern. Computer processing power
continues to rise at a remarkable clip,13 sophisticated
analysis and pattern recognition is at best in the middle of
its development arc14 and artificial intelligence applications
remain at a very early stage of development. As such,
computers should remain a productivity accelerant for the
foreseeable future.
On the other hand, the specific ability to communicate more
easily with others – via email or video conferencing, say –
seems less revolutionary. To be sure, these continue to
advance productivity, but they are less foundational than
the development of the telephone or the telegraph before it,
inventions that first made distant communication possible in
something approximating real time.
Networks
Robots
Built atop a backbone of computers, network technology
has advanced rapidly in recent decades, and the
transmission of information electronically represents
another general-purpose technology that will pay dividends
for decades and is already enabling additional innovations
that sit on top of it.
Some argue that the provision of Internet access to homes is
a far less monumental development than the arrival of other
general-purpose technologies like electricity or plumbing.
Perhaps this is true from a casual, day-to-day perspective.
But the ability to obtain virtually any information (and,
increasingly, products) over the internet – from libraries’
worth of books to entertainment to a local restaurant
review to a competitive price for a product – remains an
enormous leap beyond the glacial advances of knowledge
transmission over the past century, and is arguably on par
with the invention of the printing press and the diffusion
of widespread literacy. Decision-making is now much more
informed.
The significance of smartphones is debatable. In a sense
they are just smaller computers, and as such welcome but
unrevolutionary from a productivity standpoint. However,
their mobility is important for location-specific applications
such as navigation, and in many emerging economies they
provide the only form of computing power and Internet
access. This is truly revolutionary. Going forward, the
smartphone offers tantalizing potential as it is better
13
In line with Moore’s Law of processing power doubling every two years, and
also because computer algorithms have been rapidly improving (with quantum
computing nearing viability).
Some strategies such as machine learning have the potential to be especially
potent.
14
The so-called “internet of things,” in which even small
objects connect to the network, will be a boon for logistics,
permitting even more efficient just-in-time inventory systems
and automatic product replenishment.
Much as computers took several decades to fully permeate
society, increasingly sophisticated robotic technology now
exists and is becoming more relevant. Defined simply, robots
are computers plus a physical presence that interacts with
the world. Robots are another “general purpose technology”
in the sense that they can be deployed for a wide range
of purposes, most obviously in manufacturing, but also in
transportation and areas as far afield as personal care.
Robotic efficiency is now improving by about 5% per year
thanks to a mix of improved brains, better brawn and greater
flexibility, while costs are falling by 2% to 3% annually.15
Relative to humans, whose real wages are rising roughly in
line with their productivity, this amounts to a robot-human
cost tradeoff that favours robots by an additional 7% to
8% per year. We discuss the outlook for automation and its
potentially profoundly negative implications for employment
in Appendix D.
The scope for further automation via robots depends on
the industry. Car manufacturing is already 80% automated,
whereas many other manufacturing sectors are still very
low in their degree of automation and thus ripe for change.
The initial slow uptake is in many cases due to greater
product customization or rapidly changing specs that have
historically hindered robots. But this may cease to be such
an impediment in the future, and robot sales are now rising
by almost 30% per year.
Robots are already quite adept at tasks demanding strength,
endurance and precision. The main advances going forward
relate to manipulation (being more dexterous), perception
(having a better awareness of the world around them) and
From Bank of America Merrill Lynch, “Robot Revolution”, November 3rd,
2015.
15
| 13
task flexibility (being able to do more than one thing). More
distant is a material improvement in goal-setting (identifying
what needs to be done in a complex environment) and social
intelligence (engaging with humans and human emotions in
a pleasant, responsive and ultimately useful way).
Transportation
The remainder of the technologies we discuss are important
but not quite general-purpose technologies. As such, they
can help productivity growth but are unlikely to sustain it for
decades, or to unleash significant additional efficiencies in
their wake.
As noted earlier, it is tempting to imagine that transportation
technology has stagnated for many years. Superficially, it
has, in that airplanes are no faster than they were decades
ago and cars are stuck in more gridlock than ever. However,
tempering this initial dour interpretation, a flight is now
much cheaper than in decades past, healthier (there’s no
smoking on board), there are far more flights to choose from
and planes are much safer. These are all forms of improved
quality and thus productivity. Meanwhile, car quality has
increased substantially in the sense of being safer, more
reliable and less polluting.
Without question, the most important new technology in
the transportation space – and arguably the most exciting
development in all of technology to the lay person – is the
expected arrival of the driverless car. This achievement is of
course a manifestation of other more basic revolutions in
computing, networks and robotics.
Google has long had several autonomous cars patrolling
Mountain View, California, with millions of miles on their
collective odometers. Traditional car companies have begun
testing their own technologies. Other tech firms, including
Apple, Uber and Tesla, have also entered the driverless-car
race.
There are two drastically different approaches being
taken. The first strategy is to build a fully-automated selfdriving vehicle from scratch, a path that Google is pursuing
with particular enthusiasm. The alternate strategy is to
incrementally increase the automated features of presentday cars until they eventually cease to require human
assistance. Traditional car companies have been inching
along this second route for decades, starting long ago
with cruise control and proceeding much more recently to
automated parallel parking, automatic emergency stopping
and improved sensors that warn of nearby danger. Tesla’s
latest cars can now navigate autonomously within a lane,
remaining below the speed limit and at a safe distance to the
14
|
Economic Compass • Issue 37
car in front of them. They can also safely switch lanes when
instructed to do so by the driver.
While these advances are impressive, most estimates put a
fully driverless car at five to 15 years into the future. For all of
its much-ballyhooed miles, the Google car still cannot handle
unfamiliar roads very well, or adverse weather at all. Chaotic
traffic situations are a challenge. These are difficult nuts to
crack.
That said, once the technology is available, it could spread
surprisingly quickly. The main challenge for self-driving cars
is in honing algorithms, not the quality of the hardware.
This is important, as software can be updated for vehicles
already on the road, with Tesla’s automated features rolling
out practically overnight to existing drivers. The incremental
improvement strategy has another huge advantage: car
companies can monitor how their technology is working in
millions of vehicles and use that copious flow of information
to refine future generations of the technology. As things
stand now, Google is stuck processing data from a few dozen
vehicles.
Driverless cars may unleash a variety of fascinating
productivity-enhancing changes on the world:
nnCommuting times and traffic could be substantially lessened
by vehicles taking the best possible route, reducing the gap
between cars, driving at an efficient and steady clip, and via
the use of coordinated convoys.16
nnAlternately, it could become more practical to live a greater
distance from work, providing more affordable housing
options or a more desirable setting for families.
nnReduced traffic fatalities will prevent the loss of many
workers in the prime of their lives.
nnAdditional leisure time while traveling by car.
nnGreater transportation freedom for children and
non-driving seniors.
nnParking could be relegated away from prime real estate,
freeing up central land for more productive uses.
Despite these very real implications, the feverish excitement
over the driverless car may be somewhat overdone. The
traffic potential is real, but to the extent that it increases
people’s tolerance for driving (and appetite for distant
homes), the result may just be additional distance covered
per vehicle and thus the same old traffic snarls. The
additional leisure for the average person may also be less
exciting than it first seems. Keep in mind that three out of
16
Coordinated convoys would also allow better cooperation with smart traffic
signals.
Economic Compass
the four people in a car already enjoy this “leisure,” and
generally do not revel in book reading, sleeping or workrelated tasks. Limited mobility within the vehicle and motion
sickness will remain key impediments.
Meanwhile, renewable energy technologies – most obviously
wind and solar – continue to advance. Solar in particular has
sustained a rapid rate of productivity growth, and appears
set to be cost-competitive with oil within a few years.
The right way to frame this technology is in the context of
whether it is more revolutionary than past transportation
breakthroughs. The answer is almost certainly “no.” The
invention of the railroad radically altered the ability of
goods and people to travel long distances. The invention
of the car gave unprecedented freedom to individuals and
enabled the creation of suburbs. In contrast, the self-driving
car is nice and represents an important contribution to the
advancement of transportation technologies, but may not
outright accelerate it.
The knock on alternative energy is that it is unreliable given
the fickle nature of the sun and the wind. But when solar and
wind technologies are paired with further advances in battery
technology, this could provide a tipping point for alternative
energy production.17 Improved storage technologies – if
possible on a grand scale18 – would also materially reduce
the need for the excess capacity on the electrical grid that is
only required on the hottest few days of the year. As such, we
view battery technologies as highly significant, and with the
potential to accelerate productivity.
Drone technology is a related transportation innovation. It
may yet enable the near-instantaneous delivery of goods
as per the vision of retailer Amazon, but this isn’t obviously
a huge leap forward from the one-day service offered by
traditional package delivery companies, and there are many
safety and practical considerations that have yet to be
solved. From a policing and (especially) military perspective,
on the other hand, the applications for drones seem highly
significant. This technology could increasingly redirect
military personnel into other more productive pursuits.
Emerging markets stand to gain in particular from the
interplay of alternative energies and battery technologies,
as alternative energy sources can be deployed in a localized
way without massive infrastructure investment.
Materials
There continues to be steady progress in the materials
science space, with the development of new composites
and the recent manufacturing of graphene – an incredibly
strong and light carbon substance that is beginning to find a
market. Incremental progress continues on many fronts in the
search for materials exhibiting unusual strength, conductivity
and/or flexibility.
However, compared to the groundbreaking materials science
advances of the 20th century – with the invention of plastics
at the top of the list – our sense is that new materials
technologies are providing less support to productivity
growth than in the past.
Energy
Energy technologies are becoming increasingly interesting,
with recent innovations set to drive productivity growth in the
sector at least as quickly as prior innovations did over the
last century.
In the oil and gas space, shale technologies continue to
advance particularly quickly, possibly even spurred on by
the oil shock. The technology still has ample opportunity to
spread more broadly outside of the U.S.
Lastly, biofuels continue to hold promise as the search
continues for genetically engineered organisms that might
eventually produce fuel with little waste. Such technological
pursuits are still in the early stages.
Food
Food productivity has increased to an incredible degree over
the past few centuries. We believe farm yields can continue
to rise at a similar rate.
The strategy is no longer primarily about mechanization since
farms are now highly mechanized and have only limited
reliance on humans. The new opportunity is in optimizing
farming practices and improving the quality of produce.
Farm analytics seem to be advancing quite quickly as more
rigorous record-keeping, better and more sensors, pattern
recognition technologies, drones and driverless tractors
combine to ensure that crops are farmed in the most effective
way possible.
Similarly, the genetic engineering of crops is still in its very
early stages. Much as corn began its domestication journey
10,000 years ago as a tiny ear with rock-hard kernels before
ending up today as a juicy product that is 40 times larger,
there remains ample room for improvement right across
the spectrum of foods with regard to their ease of growth,
disease resistance, nutrition and environmental footprint.
17
Incidentally, improved battery technologies are also a key enabler for
widespread adoption of the electric car.
18
Conceivably, electric cars could provide this large-scale storage capacity.
| 15
Health
Health improvements have been enormous over the past
century, thanks to a mixture of better medical practices,
better diets, better living conditions and improved drugs.
Alas, the rate of drug development seems to have slowed
for infectious diseases, but it remains normal for other
afflictions.
Elsewhere, the gains and prospective gains are quite
impressive. We see important developments in health
tracking and diagnosis. The former is aided by a myriad of
increasingly sophisticated labs-on-a-chip, genetic tests,
electronic recordkeeping and personal health-measurement
devices. Better diagnostics are largely the result of more
informed patients thanks to the internet, and more upto-date doctors due to online medical journals. Artificialintelligence software is also becoming better at interpreting
symptoms and asking the right questions. Robotic surgery
continues to progress, with machines assisting in tasks for
which precision is key.
Synthetic biology – manipulation of the genetic code – still
holds a great deal of promise: basically, fixing problems in
the genetic code or creating new organs in a lab. “Big data”
analytics will likely help in identifying the relevant genes.
Analytics will also be key in processing the rising tide of
information coming from personal health trackers into
a usable form so as to identify the specific health risks
of individuals, and also to reach broad epidemiological
conclusions for the benefits of humankind.
TEXTBOX A: THE VARIED IMPACT OF TECH CHANGE
In fairness, not all technological breakthroughs have the
same effect on productivity (Exhibit 26):
nnThe best type of new technologies enhance the utility
of existing technologies. Examples include Internet
search engines and the smartphone, both of which
have greatly extended the value of the Internet. The
good news is that we are seeing more of this sort of
technology.
nnThe least helpful type (though still, to be clear, a
positive for productivity growth) are innovations that
displace existing technologies. Examples include
Uber’s potential replacement of taxis or financial
technologies that might displace banks. The less
positive news is that we are seeing more of this
displacing type of innovation.
nnFinally, we are seeing fewer innovations of the sort that
neither displace nor enhance existing technologies.
Examples include the invention of antibiotics or the air
conditioner.
On balance, the first two of these forces largely
neutralize one another, leaving a fairly normal mapping
of innovation onto aggregate productivity growth.
Exhibit 26: Innovation types
Customization
This category is a grab-bag for technologies that allow
for ever-more customized experiences and products. To
the extent that these better meet their buyers’ needs,
they represent an increase in effective quality and thus
productivity.
Custom manufacturing is already reasonably well advanced.
The interplay of the Internet, more sophisticated robotics
and superior supply-chain management has allowed for the
blossoming of mini-industries around customized machine
parts, personalized running shoes, book-form family photo
albums and even self-published novels. 3D printing is still
in its early stages, but promises to make some of this much
easier.19
19
3D printing should also allow the creation of more sophisticated products
not currently possible with standard molding techniques, and stronger, lighter
structures. Waste is also reduced. Applications have even started to include
components for the human body.
16
|
Economic Compass • Issue 35
Least helpful for
� productivity growth
� incumbent firms
Most helpful for
� productivity growth
� incumbent firms
Shifting
composition:
Description:
Innovation that
displaces existing
technologies
Innovation that
doesn’t affect existing
technologies
� Antibiotics
Example: � Uber
� Financial technology � Air conditioner
Source: RBC GAM
Innovation that
complements existing
technologies
� Internet search
� Smartphone
Economic Compass
Lastly, the realm of “niche targeting” has exploded thanks
to the internet, allowing individuals to organize and find
like-minded communities that simply weren’t practical
given the geographic constraints of the past. This has led to
a boom in practically every kind of hobby, a golden age of
entertainment choices, more specialized education degrees,
and even (theoretically) better marriage matches!20 From
a business perspective, companies can much more easily
identify and target prospective customers, and can also
afford to produce highly-specialized products that were not
previously practical.
Exhibit 27: EM compositional effect keeps global
productivity rising
Productivity growth (ppt)
9
Other productivity considerations
Finally, we step back from specific technologies to consider
additional productivity-relevant developments. Broadly, we
find there to be slightly more negative ancillary influences
than positive ones, though it is hardly a one-sided affair and
the collective drag should be insufficient to interfere with
our overarching conclusion of a normalization of productivity
growth.
Positive forces
We start with the positive forces. First, many national
governments are responding to the post-crisis economic
malaise and now beginning to deliver productivity-enhancing
structural reforms such as reducing the rigidity of their labour
markets, cutting red tape, and allowing financial markets
(and indeed market forces more generally) to operate more
freely. A significant swath of emerging-market countries,
Japan and peripheral Europe are implementing this strategy.
Second, it is worth remembering that many of the
productivity tailwinds that are sputtering in the developed
3
0
RBC GAM projections
-3
1992
New technology implications
Our conclusion – assisted by a brief aside into the
different styles of innovation in Textbox A – is that the
rate of technological innovation should be somewhat
faster than normal in the future. The key supports for this
argument are the still-accelerating benefits from generalpurpose technologies such as computers and networks,
the emergence of robotic technologies, and important
upticks in energy and customization. We discuss the sectorspecific implications of these new technologies in Appendix
C. One slight tempering factor from the perspective of
overall productivity growth is that the rapid advance of new
technologies can actually cause the value of existing capital
and labour to depreciate more quickly.
6
2001
EM
DM
2010
Global
2019
2028
Note: Productivity growth includes capital, labour quality and total factor
productivity growth. Global is an aggregate of six EM countries and thirteen
major developed countries. RBC GAM estimates used in projections for
2015 and beyond. Source: Feenstra, Inklaar and Timmer (2015), “The Next
Generation of the Penn World Table” forthcoming in American Economic
Review, available for download at www.ggdc.net/pwt, The Conference Board
Total Economy Database, RBC GAM
world are still blowing freely for emerging economies.
These include the rapid pace of urbanization, rising female
employment, improving education levels and rising human
longevity. Furthermore, these countries can still expect to
benefit from falling corruption and better infrastructure over
time.
Third, and continuing with the emerging-market theme,
there is a favourable compositional effect at work. Emerging
economies represent an ever-rising share of global output,
and their rate of productivity growth is faster than that of
developed nations. This allows global productivity growth
to – in principle – accelerate even if national-level
productivity growth remains unchanged (Exhibit 27).
Fourth, across the long arc of history there is tentative
evidence of a “challenge-response mechanism” that
surfaces during periods of weak population growth whereby
productivity growth ticks higher in a compensatory manner.
The poor demographic outlook for the next few decades
could yet unleash this force.21
Fifth, a reduction of “too-big-to-fail” subsidies for financial
institutions has resulted in lower relative profitability for
banks and other financial companies, impacting worker
compensation. If this trend persists, it might improve
the distribution of talent across a wide range of sectors,
improving overall productivity growth.
21
The internet permits the vetting of a far larger swath of people for dating
purposes, theoretically improving the quality of marriage matches, and thus
happiness.
20
The causality is not entirely understood, but an inadequate labour supply
could stoke the embers of innovation to increase the output of the existing
workforce.
| 17
Negative forces
However, there are also a number of ancillary negative forces
to consider on productivity.
First is demographics. An aging population generally
increases skill levels in the workforce since older employees,
particularly white-collar workers, are materially more
experienced and knowledgeable than their younger
counterparts. We calculate that this structurally adds around
0.2 percentage point per year to U.S. productivity growth.
However, we cannot simply append this to our productivitygrowth forecast because the benefit of a rising workforce age
was even greater over the past few decades. Thus, this is a
tailwind, but a fading one.
Second, and continuing with the demographic theme, there
are clear productivity negatives associated with an aging
population. One is that while experience levels rise, the rate
of innovation plummets for workers past middle age. Another
is that governments are forced to expand their entitlement
spending, foregoing other more productivity-enhancing
purposes.
Third, and in direct opposition to the hope of additional
structural reforms, there has been a worrying trend toward
more populist governments in the developed world –
ones that are less rather than more inclined to implement
productivity-enhancing reforms.
Fourth, the world is transitioning from U.S. hegemony to
a multi-polar world given China’s economic ascendance.
Historically, multi-polar environments have been associated
with less globalization, and by extension less economic
growth and smaller productivity gains.
Fifth, rising inequality is known to limit productivity growth
as it constitutes a less efficient allocation of resources. While
it is important to allow market forces and meritocracy to
operate at a fundamental level, it is equally undeniable that
the poor have a higher marginal propensity to consume their
money, derive a larger benefit from each additional dollar of
income and enjoy superior returns to education and health
spending. Productivity suffers when the distribution becomes
too lop-sided.
Sixth, the prospect of further climate change is a notable
negative for productivity growth, with a two degree Celsius
increase in global temperatures expected to subtract around
4% from global productivity.
18
|
Economic Compass • Issue 37
Seventh, because monetary policy and fiscal policy have
both been significantly stretched by the most recent crisis,
there will be diminished policy recourse in any future
downturn. Thus, economic downturns could be more
frequent and/or deeper in the future.
Outlook
The outlook for productivity is hugely important, both in its
role as the central enabler of rising financial well-being and
in the context of deteriorating demographics as the only
means of avoiding economic stagnation.
Our forecast is as follows. Over the next few years,
productivity growth should remain underwhelming as the
cyclical depressants of the financial crisis continue to play
out.
However, over the long term, we see no reason why
productivity growth cannot revert to a normal rate. The yoke
of cyclical depressants should fade with time. Granted,
capital investment and labour quality are likely to provide
slightly less help than normal, and the diffusion of existing
technologies around the world may even fade to some
extent. But we see considerable potential in a raft of powerful
new technologies, and so believe that innovation – helped
by a rising contribution from emerging economies – should
accelerate in the future. This latter force should roughly offset
the drags from capital and labour quality, permitting a return
in both the developed and emerging world to the sort of
productivity gains that prevailed prior to the distortions of
the past decade.
For the developed world, this should amount to around
1.75% productivity growth per year (Exhibit 28). For emerging
economies, we figure this should average 4.65% (Exhibit
29). Naturally, the actual experience will vary significantly
over time and by country, and emerging-market productivity
growth should ebb over the long run as these nations
become wealthier.
For businesses and investors, this relatively optimistic
conclusion is a relief. Yes, investment returns could be
somewhat lower in the future for demographic reasons,
but the core foundation of corporate growth – businesses
building a better mousetrap and finding new markets –
should remain active.
That said, from the perspective of individual businesses,
the future is still fraught with risk. Technology is radically
Economic Compass
changing how many economic sectors operate. Myriad
existing businesses will be disrupted by new entrants.
The scalability of new business models is such that
economies are now increasingly oriented toward winnertake-all outcomes, with the implication that picking the right
company will be crucial.
2.2
1.99
2.0
0.09
1.8
1.6
1.4
1.2
TFP
0.4
1.75
0.14
0.44
0.41
0.97
0.04
0.8
0.6
7
0.36
1.0
0.01
0.30
0.22
1.10
0.90
0.70
0.2
0.0
1980-2003
Tech diffusion
2004-2014
Tech innovation
Labour quality
Forecast
Capital
Note: Contribution to GDP growth of thirteen major developed countries.
Total factor productivity (TFP) is separated into innovation and diffusion
of technology. Forecast based on RBC GAM estimates. Source: Feenstra,
Inklaar and Timmer (2015), “The Next Generation of the Penn World Table”
forthcoming in American Economic Review, available for download at www.
ggdc.net/pwt, The Conference Board Total Economy Database, RBC GAM
Exhibit 29: Emerging markets: tech innovation to ramp up
Contribution to GDP growth (ppt)
Contribution to GDP growth (ppt)
Exhibit 28: Developed nation productivity growth
should revive
For people, further productivity growth is good in the sense
of potentially raising real wages and the standard of living,
but it is a significant threat over the longer run if automation
eventually translates into mass unemployment, as discussed
in Appendix D. It will be up to voters (and human nature) as
to whether this is allowed to happen, and whether the end
result is utopia or dystopia.
6.11
6
0.86
4.43
5
4
3
TFP
0
0.50
0.71
0.75
0.40
0.47
0.65
4.38
2
1
4.65
0.57
0.30
3.00
2.60
1992-2003
Tech diffusion
2004-2014
Tech innovation
Forecast
Labour quality
Capital
Note: Contribution to GDP growth of six EM countries. Total factor productivity
(TFP) is separated into innovation and diffusion of technology. Forecast based
on RBC GAM estimates. Source: Feenstra, Inklaar and Timmer (2015), “The
Next Generation of the Penn World Table” forthcoming in American Economic
Review, available for download at www.ggdc.net/pwt,The Conference Board
Total Economy Database, RBC GAM
| 19
APPENDIX A
GDP DISTORTIONS
Depending upon how broad a net one wishes to cast,
one could further add the perfectly legal economic
output generated via unpaid household production:
your own cooking, cleaning, child care, housework and
gardening. The value of this output is no less significant
when generated personally rather than outsourced to
paid help. The collective value of unpaid household
production is regularly pegged at another 10% to 15% of
GDP in the U.S.
While these observations coherently argue that the
aggregate level of human output is at least 26%
higher than officially reported, the interpretation is
slightly different when the focus is shifted from the
level of output to the rate of growth. This is because
the underground economy and unpaid household
production have both been declining as a share of
GDP for decades. Thus, a proper accounting of these
excluded sectors actually subtracts 0.2 percentage point
per year from properly measured GDP growth, rather
than adding to it. Of course, we care about productivity
for the purposes of this report. The interpretation there is
positive, but only very slightly. Because the unreported
economic output tends to be materially less productive
than official output, a declining underground economy
means that a proper accounting of productivity growth
has been as much as 0.1 percentage point faster per
year than the official estimate over the past several
decades.
Exhibit A: U.S. after-tax corporate profits at record high
12
10
U.S. after-tax corporate profits
as % of GDP
Official GDP figures simply do not capture several
types of economic activity, collectively entitled
“the underground economy.” At the extreme end of
lawlessness, the official numbers fail to capture such
activities as illegal drug manufacturing and dealing,
prostitution and illegal gambling. Also excluded are
less malignant but nevertheless illegal businesses
engaged in tax avoidance, such the proverbial plumber
paid mainly in cash and waiters who do not report
their tips. There are then rather more benign cash-only
“businesses” such as babysitting and lemonade stands.
None of these is captured in GDP, and the omissions
across the OECD are thought to average about 16% of
economic output, and significantly more for emerging
economies.
Long-term average: 6.49%
8
6
4
2
0
1980
1985
1990
1995
2000
2005
2010
2015
Source: Haver Analytics, RBC GAM
Fortunately, this muted finding is not the end of the story.
There are a number of other arguments as to why true
productivity growth is likely faster than the official figures:
nnSurging corporate profit growth has easily outpaced
economic growth over the decades (Exhibit A).1 This
offers a hint (but nothing more) that official GDP and
productivity growth may have been undercounted in
recent years.
nnThe service sector is becoming substantially more
important over time, and is notoriously difficult to
measure. How does one even begin to measure the
economic output of a bank, for instance, let alone the
government? Statistical techniques adapt to changing
conditions with a lag, leaving elements of the service
sector perpetually under-measured.
nnRepeated studies have found that the official inflation
figures overestimate true inflation in large part by failing
to incorporate the deflationary effects of new technologies
sufficiently quickly into the price basket. Because
productivity growth is derived from real GDP, and real
GDP comes from nominal GDP less inflation, any
overestimate of inflation is also an underestimate of
productivity growth.
nnOn a related front, statisticians still fail to fully capture
the rising quality of technological products. An hour
of internet usage today offers a much higher quality of
service than it did in the past. The bandwidth is higher,
the total data downloaded will almost certainly be higher,
1
As a counterpoint, revenue growth has not outpaced GDP growth, rendering
a somewhat murky interpretation.
20
|
Economic Compass • Issue 37
Economic Compass
and the range of online content is exponentially greater.
These are unrecorded productivity gains.
nnEvery year seems to bring an ever greater kaleidoscope
of products to choose from, and a greater variety of ways
to consume it. This better selection increases the odds
of finding just the right product. Being able to consume
your favourite TV show or film with the click of a button
is surely worth something, even if just in the sense of
leisure more productively spent.
nnPrices regularly fail to capture the full economic value
of a product. Theoretically, people don’t bother to buy
something unless it is worth more to them than the
asking price. This unrecorded “consumer surplus” is often
small, but sometimes it can be profound. The surplus has
arguably grown in recent years as the “economy of free” –
a myriad of free or nearly free services – have cropped up
on the Internet, ranging from Wikipedia to online news to
social networks. These free services clearly have a value
greater than zero, and so productivity gains are not fully
captured. Even where ads or fees create an implicit cost
for the user, it is usually well below the consumer surplus.
Various studies calculate that the consumer surplus
from information technology has grown to over US$100
billion per year in the U.S., or at least 0.5% of GDP. Others
conclude that productivity growth may now be as much
as 0.2 percentage point per year faster than the official
numbers show.
nnProductivity figures capture process improvements more
easily than product innovation. It is easy to observe
that a factory is now producing more goods without
changing the number of machines or workers. It is
much harder to determine how much productivity has
risen when a factory switches from an old product to a
brand new product. Some believe product innovation
has accelerated over the past few decades, meaning
productivity figures may be underestimated.
| 21
APPENDIX B
WELL-BEING
Inequality: A country’s well-being would hardly be the
same if all of the money were in one person’s hands.
Thus, the degree of inequality matters. The increased
inequality visible within countries in recent decades
argues that well-being has not kept pace with GDP,
particularly since the poor generally enjoy a higher return
on each dollar deployed into health and education.
However, the assessment becomes less negative when
contemplated from an international perspective, as
inequality has actually declined between countries.
Sharply declining poverty rates in middle-income and
low-income countries demonstrate that even modest
income gains can sharply reduce poverty rates
(Exhibit B).
Leisure: A large fraction of leisure goes uncounted
in GDP, yet provides enormous benefit to well-being.
Surely there is value in time spent not working, be it
lazing around or engaging in hobbies. As the world’s
population ages, the retired fraction of the population
will continue to grow. Society as a whole is set to
experience a marked increase in leisure (Exhibit C), even
if the prophecies of Appendix D do not come true.
Health: Although median real income levels have
stagnated in the developed world for several decades,
health outcomes have generally improved. Longevity
continues to rise. Even in the poorest of developing
economies, infant-mortality rates have enjoyed a
significant decline despite underwhelming economic
22
|
Economic Compass • Issue 37
92
80
90
75
88
70
86
65
84
60
82
55
80
50
78
76
45
Decline in poverty accelerates
even as global growth slows
40
35
74
72
1980
1984
1988
1992
1996
Lower income nations (LHS)
2000
2004
30
2012
2008
Middle income poverty headcount ratio
(% of population)
Environment: Environmental destruction is not captured
directly by GDP, but certainly matters to well-being.
Water and air pollution has declined precipitously in
the developed world for decades, but carbon emissions
are becoming an ever-more-pressing problem. It is more
uniformly negative for emerging economies. These
collectively leave a mixed interpretation, but ultimately a
slightly negative one.
Lower income poverty headcount ratio
(% of population)
Debt: Global indebtedness rose markedly for several
decades, halting its upward ascent only recently. The
acquisition of this debt permitted a temporary boost
to GDP that, if anything, damaged well-being given the
need to eventually pay the money back.
Exhibit B: Well-being is likely improving faster than GDP
Middle income nations (RHS)
Note: Poverty threshold is set at $3.10 a day based on 2011 PPP.
Source: Haver Analytics, RBC GAM
Exhibit C: Structurally less work = more leisure
55
53
Labour force participation
(% of total population)
GDP and productivity, even when properly measured, are
not perfect arbiters of our well-being. They fail to reflect
the impact of debt, the environment, inequality, leisure
time, health and personal safety.
...more
leisure
51
49
47
Less
work...
45
43
41
39
37
35
1955
1965
1975
1985
1995
2005
2015
Source: Haver Analytics, RBC GAM
growth. These are hugely important welfare gains that are
not fully captured in GDP.
Safety: Despite perceptions to the contrary due to blanket
media coverage, the developed world is a much safer place
than it was a few decades ago. Rates of violent crimes are
almost universally down.
When we assess these collectively, we find one factor
(debt) that argues well-being is rising less quickly than
GDP, two (environment and inequality) that make a mixed
but ultimately negative assessment, one (leisure) that
makes a weak positive argument, and two that reach
a clearly positive conclusion (health and safety). Thus,
we conclude that well-being may be rising slightly more
quickly than even properly measured productivity figures
would suggest.
Economic Compass
APPENDIX C
TECH CHANGE AT THE INDUSTRY LEVEL
Technological change has already had a revolutionary
effect on the retail sector, displacing book stores and
record shops, and increasingly biting into other retail
categories (Exhibit D). This process has ample room to
continue given that just 7% of retail sales are conducted
online. Fundamentally, the allure of e-commerce is that
it enables much greater transparency, providing buyers
with the opportunity to secure the cheapest price and
offering a better sense of quality via user and expert
reviews.
The media and entertainment sectors have been
substantially undercut by free competition, the rise of
citizen journalism and pirating. Going forward, a large
swath of the service sector suddenly seems similarly ripe
for disruptive technological change.
Manufacturing may also evolve quite quickly, if more
due to robots and automation than the Internet and
algorithms that threaten other sectors.
Several additional themes warrant exploration. A
winner-take-all economy seems to be forming, at both
the worker and corporate levels. While the forces of
creative destruction have always pushed out tired old
companies in favour of innovative new ones, it seems
that a multitude of old firms are being replaced by only a
handful of new ones. The arrival of Amazon in the retail
space is a good example, and illustrates the incredible
scalability of the new breed of businesses. For investors,
this means it isn’t enough to invest in the right sector –
they must also identify precisely the right company
within it.
Within sectors, the declining power of “brand” – as
consumers are able to more easily vet actual quality
rather than rely on a company’s reputation as a proxy –
is a significant disruption.
Disintermediation – removing the middle man from
transactions – is another theme. For books and many
types of merchandise, this effectively means buying
straight from the wholesaler online, without the need
for physical shops. Do-it-yourself online property
listing will reduce the need and/or pricing power of real
estate agents. LinkedIn and a variety of job-listing sites
Exhibit D: E-commerce is becoming more pervasive
Electronics
Fashion and apparel
Services
Books
Tickets
Music and games
Home applicances
Beauty products
Home furnishings
Sports and outdoor
Toys, kids and babies
Household items
Groceries
0
20
40
60
80
100
% of respondents that shopped online in the past 3 months
Note: % is the global average of the U.S., U.K., Germany, Japan, India, Brazil,
Russia, China, South Africa and Nigeria. Source: A.T. Kearney Connected
Consumer Study, RBC GAM
reduce reliance on traditional head hunters. With peer-topeer lending, people can lend and borrow directly from one
another, without the involvement of a bank. The list goes on.
A final theme is the “sharing economy”, which captures
the idea of renting goods and services rather than buying
them outright. Burgeoning examples abound. Apartments
can be rented out on a short-term basis via services like
Airbnb, or for a longer period via other services. Cars (and
drivers) are effectively rented via Uber. It is becoming more
common to rent luxury goods rather than pay an enormous
amount for something that might be worn only a few times.
Depending on how the self-driving car evolves, one line
of thought is that there could be a wholesale shift from
vehicle ownership to renting once transportation can be
summoned for any purpose and duration with the click
of a button. Fundamentally, why own rather than rent a
lawn mower or a tool when it is only used for a few hours a
year? At the most existential level, it is possible to “rent”
a worker for a specific task via a variety of websites. Most
of these pursuits were not practical before the Internet
made matching transaction partners so easy. These are all
opportunities for people and entrepreneurs, but challenges
for existing companies in each sector.
Exhibit E reviews the constellation of changes afoot in a
wide range of sectors.
| 23
Exhibit E: Tech change at the industry level
Sector
Change
Size of
sector pivot
Effect on
incumbents
Opportunity
Retail
Internet shopping increases transparency
of prices and quality; provides new
competition; offers scope for niche
products; sharing economy
Massive
Displaces many incumbents; empties Scalability and network effects
malls; lowers margins
create only a few big winners; others
must focus on niche success
News
Many new online media sources; new
advertising methods; rising variety of
content; niche targeting; low price or costfree
Massive
Displaces many incumbents;
drastically lower margins
Provide specialized content (niche
subject matter or geography)
Entertainment
Streaming of entertainment; rising content
quality; niche targeting; new pricing
models; self-publishing; pirating
Massive
Displaces many incumbents;
eventually more profits to go around
Spectacle of live theatre/concerts;
streaming business; online ticket
sales; niche focus
Advertising
Online advertising captures market share,
allows incredible targeting
Massive
Google has stranglehold on online
advertising; incumbents are left with
rest
Online advertising; niche advertising
Military
Drone technology; cyber warfare; selfnavigating vehicles
Massive
May revolutionize warfare
Rapid change in military
expenditures
Massive
Displacement of taxi medallion
Fractured industry to become more
owners; eventual displacement of taxi concentrated; package delivery
drivers and truck drivers
Transportation Demand-sensitive ride-hailing pricing,
better quality assurance; self-driving
services
car; self-driving transport truck; package
delivery
Information
technology
I.T. driving productivity gains; entering
many other sectors
Massive
Big opportunity, but smaller nimble
firms also threaten when network
effects/scalability have not locked
down category
Continue disrupting other sectors
Education
Improved testing; improved teacher
analytics; Internet and video conferencing
should allow all to be taught by best
teacher & curriculum – improves quality
and offers incredible scalability
Massive
Price increases to slow; brand name
quality premium to persist, lower
prestige schools to struggle
Invest in technologies that improve
quality of testing, student/teacher
analytics, scalable education
solutions
Massive
Cost savings; developed-world
manufacturers compete more equally
with emerging markets
Invest in robotics; developed-world
manufacturing
Moderate
Incumbents to be more reliant on
acquisitions of new medicines from
smaller firms
New, smaller companies to generate
many of the new technologies and
drugs
Motor vehicles New competitors (tech + Tesla); electric car;
self-driving car; sharing economy; possible
undercutting of dealer model
Moderate
Incumbents can survive but there
might be reduced to mere assembler
for tech firms
Keep pace with tech innovation in
sector; disruption to dealer sales
model
Finance
Non-banks skirt regulations; robo-lending;
ETFs; online functionality; peer-to-peer
lending; new payment systems; big data;
open source; heavier capital requirements;
tighter bank regulations
Moderate
Sector function essentially
unchanged, but with new
technologies to improve/speed
processes, new profitability
headwinds
New entrants provide real
competition in wealth management
and payments; but banks have trust
of clients and will largely acquire
new technologies
Energy
New types of oil and gas production; rapid
tech improvement in renewables; electric
car; battery technologies
Moderate
Traditional energy production
still cost effective; but electric car
threatens type of energy needs
Further investment needed in
renewables, battery tech; potentially
new pricing model
Law
Low-cost plain-vanilla law services; AI
assistance in research
Moderate
Displaces small law firms; AI helps
sophisticated firms
Low cost plain vanilla law services
Utilities
Low oil and gas prices; renewables boom;
improving battery technology; electric car
Minimal
Lower peak demand for power with
better battery tech; modest threat
from non-utility grid-connected
renewables; but opportunity with
electric car
Expand renewables and natural
gas; invest in battery tech; electric
charging stations
Telecom
Smartphone revolution; increased data
usage; less voice usage; lower voice prices
Minimal
Incumbents largely remain, with shift
in sources of profitability
Focus on data; further EM growth
Hospitality
Peer-to-peer property rentals; sharing
economy
Minimal
Does not seriously displace hotels;
Restaurants and hotels seem
potentially greater restaurant demand durable amid tech change
Manufacturing Automation to continue reducing costs and
lowering reliance on labour
Health
Self-diagnosis on internet; robo-surgeries;
lab-on-a-chip; health tracking; computer
diagnosis; electronic record-keeping;
synthetic biology; analytics
Source: RBC GAM
24
|
Economic Compass • Issue 37
Economic Compass
APPENDIX D
AUTOMATION AND EMPLOYMENT
Manufacturing employment has suffered in more recent
decades, but once again there has been no material
surge in unemployment as these workers were absorbed
elsewhere.
However, the story may not end in quite so cozy a
fashion. The world is now changing in rapid and
unprecedented ways. The long-derided threat of
technology-induced job losses on an economy-wide scale
may actually be on the cusp of coming true. The correct
analogy may not be the successful repurposing of farm
workers but the absolute obsolescence of workhorses
after the introduction of motor vehicles in the early 20th
century.
One initial symptom of a profound shift is that real wages
have now been underperforming productivity growth for
some time. The gains are accruing disproportionately
to the owners of capital, and less so to the operators of
capital. Some estimates attribute as much as half of the
wage underperformance to the effects of technology.
Automation is still in a nascent state, and seems capable
of stretching significantly further. Robots and computers
can already perform many tasks at a higher level than
people, and their capabilities continue to grow by leaps
and bounds (Exhibit G). At the same time, robot costs
are falling. Human abilities, in comparison, are relatively
static. As noted earlier, robot cost-competitiveness is
outpacing humans by an additional 7% to 8% each year.
It is one thing when robots can perform a few tasks better
than people, but something very different when they
start to do many things better – and more cheaply. The
next set of technological advances seem set to bleed into
occupations that were previously shielded from these
forces, such as clerks, cashiers, truck and taxi drivers,
warehouse workers and soldiers. Given all of this, it
does not require a great leap of the imagination to arrive
at projections of large scale job losses. A widely-cited
Exhibit F: Falling U.S. farm employment did not spur mass
unemployment
90
80
U.S. agricultural employment
(% of labour force)
A perpetual fear associated with technological change
– the so-called “Luddite fallacy” – is that it will result
in massive job losses as machines replace people.
Fortunately, this fear has repeatedly gone unrealized.
Increased farm mechanization and efficiency has
undeniably displaced millions of agricultural workers
(Exhibit F), yet unemployment failed to rise as other
sectors happily absorbed the excess labour.
70
60
50
40
30
20
10
0
1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020
Source: NBER, Federal Reserve Bank of St. Louis
Exhibit G: Humans versus robots
Human advantage
Robot advantage
Creativity
Task
flexibility
Social
intelligence
Goal setting
Perceiving
Manipulation
Pattern
recognition
Endurance
Strength
Precision
Calculation
Note: Arrows indicate shifting relative advantage. Source: RBC GAM
Oxford University study estimates that 47% of U.S. jobs are at
a high risk of being automated away within the next 20 years
(Exhibit H). This risk may take longer to manifest in emerging
economies, but is ultimately no less significant.
Let us be clear, however, that unemployment rates would
not rise as high as 50% even if half of all existing jobs were
destroyed since there are always new sectors being formed
with their own labour needs.
Moreover, many jobs should prove resistant to
automation given the requirement of task flexibility – for
instance, a waiter must be able to make a personalized
recommendation, take an order, deliver the food, open a
wine bottle, replenish the bread, mop up the spilled soup
and make change.
Exhibit H: Almost half of U.S. employment may be
replaced by machines
Jobs less
susceptible
53%
Humans are also still materially more creative than
robots, have a higher social intelligence and are far better
at goal setting. As a NASA report once famously quipped,
“man is the lowest-cost, 150-pound, nonlinear, all-purpose
computer system which can be mass-produced by
unskilled labor.”
But even with all of these qualifiers, it still seems quite
possible that widespread job destruction will eventually
result from this automation trend.1 Providing an indirect
confirmation of this assessment, the majority of the
companies floating to the top of this new technological
world do not have large staffs.
Businesses may view this automation trend as a source of
profit growth for now, but they risk eventually running out
of customers. As such, automation and the potential for
widespread job losses presents a challenge to all parties.
Policy remedies
What can be done? Unlike horses, people vote, and in
extreme scenarios, they revolt. It is therefore unlikely that
an underclass of mass unemployed will be allowed to rot.
Society will have to (try to) do something about it.
While many approaches may be taken, the most obvious
and conventional strategy is large-scale government
redistribution of money to the structurally unemployed,
financed via substantially higher taxes on the owners of
capital who benefit from automation, and perhaps also
in the form of higher taxes on high-income workers who
benefit from their collaboration with machines.2
Another potential policy strategy would be to pursue
large-scale “make work” projects, be they infrastructureoriented or government-funded services (more social
workers, street cleaners, park gardeners and tourist
greeters).
1
Furthermore, wage rigidities – the difficulty companies have in reducing
worker wages – may result in greater unemployment when a lower wage
might otherwise have enabled workers to remain competitive with
machines.
This might hurt productivity growth given the disincentive of higher taxes on
productive investments.
2
26
|
Economic Compass • Issue 37
Jobs
at risk
47%
Note: Jobs at risk are U.S. jobs that the authors expect could be automated
over the next 20 years. Source: Frey and Osborne (2013). “The Future of
Employment: How Susceptible Are Jobs to Computerisation?”, Oxford Martin
School Working Paper, RBC GAM
Further strategies might include banning technologies
that prove particularly disruptive to employment – though
this is a slippery slope and without question productivitydeterring – or incenting the development of technologies
that support workers rather than replace them.
Optimistically, human ingenuity may yet find a creative
solution that we cannot yet fathom.
Scenarios
The best-case scenario in an automated future is a world
that has managed to largely shed the drudgery of work,
permitting more leisure time, a greater focus on family
and friends, and enabling a world of hobbyists, artists,
philosophers and inventors, free to pursue their passions.
In contrast, there are two very bad scenarios. The first is
a world that proves incapable of rescuing the structurally
unemployed, resulting in an irreparable fracture to
society and widespread poverty. The second is a world in
which financial transfers prevent destitution, but decay,
delinquency and malaise nevertheless set in without the
structure and purpose of work.
Economic Compass
This report has been provided by RBC Global Asset Management Inc. (RBC GAM Inc.) for informational purposes only and may not be reproduced,
distributed or published without the written consent of RBC GAM Inc. In the United States, this report is provided by RBC Global Asset Management
(U.S.) Inc., a federally registered investment adviser founded in 1983. In Europe and the Middle East, this report is provided by
RBC Global Asset Management (UK) Limited, which is authorized and regulated by the Financial Conduct Authority. RBC Global Asset Management
(RBC GAM) is the asset management division of Royal Bank of Canada (RBC) which includes RBC Global Asset Management Inc., RBC Global Asset
Management (U.S.) Inc., RBC Global Asset Management (UK) Limited, RBC Alternative Asset Management Inc., and BlueBay Asset Management LLP,
which are separate, but affiliated corporate entities.
This report is not intended to provide legal, accounting, tax, investment, financial or other advice and such information should not be relied upon
for providing such advice. RBC GAM takes reasonable steps to provide up-to-date, accurate and reliable information, and believes the information
to be so when printed. Due to the possibility of human and mechanical error as well as other factors, including but not limited to technical or other
inaccuracies or typographical errors or omissions, RBC GAM is not responsible for any errors or omissions contained herein. RBC GAM reserves the
right at any time and without notice to change, amend or cease publication of the information.
Any investment and economic outlook information contained in this report has been compiled by RBC GAM from various sources. Information
obtained from third parties is believed to be reliable, but no representation or warranty, express or implied, is made by RBC GAM, its affiliates or
any other person as to its accuracy, completeness or correctness. RBC GAM and its affiliates assume no responsibility for any errors or omissions.
All opinions and estimates contained in this report constitute our judgment as of the indicated date of the information, are subject to change
without notice and are provided in good faith but without legal responsibility. To the full extent permitted by law, neither RBC GAM nor any of its
affiliates nor any other person accepts any liability whatsoever for any direct or consequential loss arising from any use of the outlook information
contained herein. Interest rates and market conditions are subject to change.
A Note on Forward-Looking Statements
This report may contain forward-looking statements about future performance, strategies or prospects, and possible future action. The
words “may,” “could,” “should,” “would,” “suspect,” “outlook,” “believe,” “plan,” “anticipate,” “estimate,” “expect,” “intend,” “forecast,”
“objective” and similar expressions are intended to identify forward-looking statements. Forward-looking statements are not guarantees of
future performance. Forward-looking statements involve inherent risks and uncertainties about general economic factors, so it is possible that
predictions, forecasts, projections and other forward-looking statements will not be achieved. We caution you not to place undue reliance on
these statements as a number of important factors could cause actual events or results to differ materially from those expressed or implied in
any forward-looking statement made. These factors include, but are not limited to, general economic, political and market factors in Canada, the
United States and internationally, interest and foreign exchange rates, global equity and capital markets, business competition, technological
changes, changes in laws and regulations, judicial or regulatory judgments, legal proceedings and catastrophic events. The above list of important
factors that may affect future results is not exhaustive. Before making any investment decisions, we encourage you to consider these and other
factors carefully. All opinions contained in forward-looking statements are subject to change without notice and are provided in good faith but
without legal responsibility.
® / TM Trademark(s) of Royal Bank of Canada. Used under licence.
© RBC Global Asset Management Inc. 2016
EC (37/2016) E
Economic Compass - Issue 37_EN 01/04/2016